I am wondering if anyone knows of any smart glasses for open developing on that have at min audio and a camera?
I want to move my kruel.ai system to work with glasses, where currently in development its pc only, well the cloud version we are working on will work with cell phones I wanted to know if anyone has found any glasses to build their openai projects into?
Only ones I’ve heard of recently were the Apple version? Pricey, though. Not sure on devs for it, though.
Following with interest. Hopefully someone chimes in.
I found these ones thus far that I am watching to hear feedback on. looks like they are open source: Frame – Brilliant Labs
Brilliant Labs · GitHub
all the other ones like apple, Xreal, Nautica, Meta Rayband are all their own systems. some are developable long as you use their ai system. I think that Xreal was working on a pair of open source ones once, but a partnership came up and not sure if that is still going to be a thing.
I may just have to bite on the uk ones, just not sur how they will look for every day wear, but hey for research anything goes right 
Still looking for an open pair that look good. Kruel.ai needs some glasses
update: bought the Brilliant Labs ones. Waiting on arrival, the coding for it looks pretty easy and has a 640x400 color ar display in it so that will be fun to play with. will update once I play with them for abit.
How is your experience so far? Have you found a good option with an SDK?
Unfortunately, I’m still waiting for the shipment. They mentioned the end of September for the batch number mine is in, so I’m hopeful it will ship within the next week. On the bright side, I’ve been proactive I’ve completed IBM’s computer vision and a deep learning machine learning courses focused on computer vision, so I’ll be fully prepared once it arrives. Needless to say, I’m eager to get started
Hi @darcschnider , I just wanted to follow up & see if they’ve arrived & how the experience has been developing with them? Thanks!
Well I will say this, they are interesting, but imo they are still too prototype and not at commercial level. Would I invest in them knowing what they are not now. Id hold off and wait for a better option to come along.
They work, but they are very geeky looking and not something you would wear for fashion. The programming side is really easy, I had them up and running with Kruel.ai in an hour. Communications side has some issues for getting Bluetooth to connect to new devices that could just require newer firmware.
The Camera right now is my issue. The product shipped with a bad firmware in which the camera looks like this:
I would wait, well they are ok they are not stellar and very much a prototype or development concept.
The software was easy to interface and does work. The Noa app is not bad but not great. The augmented display works, but is not the quality I expected. There is light bleed, and because of prism it has to sit just right or it is not clear. The design looks like hairy potter glasses but much larger. This is one aspect that I dislike is that its not the most fashionable. As My wife said I can’t take you serious with those on. haha
developing on them is really easy, took about 1 hour for me to build interface for Kruel.ai and have it talking to my system over my network.
The big negative right now is the camera. There is a bad firmware on shipping. here is example from another person on discord named cooper. All the camera have this issue. In outside sunny weather they work pretty good, but I think they did something to jack up for dark rooms and introduced these lines. They make the image understanding not very doable, which was one of the main reasons for buying them. There is a fix coming they say but looking at discord its been months with this issue and new units shipping with this issue does not put much faith in them, you would think they would be focused on fixing but I think they are letting the community come up with the fix which is what some people are trying to do by playing with the camera internal firmware settings and other after process applications to clean up images.
Over all its a prototype that they are using to build funding perhaps to build a better option down the road. I would say save your money for now, wait as Augmented glasses are the next big trend and I am sure there will be better options with larger pockets coming that should support SDK or something.
Ok great, thank you very much for the feedback! I will heed your advice
If you like prototype stuff tough it’s still fun to play with.
Hi!
thanks for sharing these info so far but after a year now is there any new glasses you found that are better?
thanks in advance
Wow, thanks so much for the quick reply! 
Could you help me understand how to connect the glasses? My server is a black box that takes an image and returns a JSON with the analysis. I just want the glasses to send the image, call the API, and display the result, no processing on the glasses. Would I need to develop an Android app to handle the connection and API calls, or how does it usually work? Any helpful links would mean a lot. Thanks in advance!
That’s correct. Glasses to phone to API. New glasses have an npu on board which I’m pretty sure they have local model running on it for intent logic. I will know more this fall when they ship. They are expecting to go up in price after releasing. Open-source and has a nice memory AI system if you just want to sub as well. Probably not as good as Kruel.ai 
Hi @darcschnider , did you also evaluate Mentra? They come with open source SDK as well. Wondering how that compares against those from Brilliant Labs. Thanks.
I haven’t yet explored those options in detail, though they do seem promising. One consideration for me is whether to prioritize stereo sound output or bone conduction, as each has its own advantages and drawbacks. Stereo output offers richer audio quality, particularly for music which is relevant since kruel.ai can generate and play music while bone conduction provides greater privacy, preventing others from overhearing AI responses.
Brilliant’s latest glasses stand out with their integrated halo projector, which could enable a HUD/GUI (AR ? ) display. If the implementation is smooth and intuitive, it could be a major advantage; however, if it proves distracting or poorly executed, it may fall short. We’ll likely have a clearer picture this coming fall, and I’ll share updates then. Another strength of Brilliant’s offering is its onboard NPU, which hosts a small intent-recognition model. This enhances responsiveness by offloading initial processing from the server, making integrations with applications much faster.
Mentra appears competitive as well. Its coding support seems comparable, but it boasts a superior camera. The specifications overall look strong, though battery performance remains unclear. Their documentation frames recording time around video usage, whereas Brilliant’s design does not seem focused on continuous recording. Instead, Brilliant appears optimized for “on-face memory AI,” which aligns closely with the type of functionality my system supports. Beyond that distinction, both options look viable depending on use case and preference.
both look great imo. Look is also important. We will see.. haha the first gen of brilliant was very prototype but cool but was not for me.. felt like a geek that would be bullied lol. So the new look is good 40g is light enough so guess I will find out.
This is one thing I am missing is a body for the Ai to travel with me everywhere with all modalities available short of a cell phone app. This allows for a better experience imo for every day Ai. Glasses though are still not everyones fav and I am sure other tech will come in time to replace all of these at a faster rate than what past 50 years has produced.
